摘要
列车初始晚点严重程度和运行图冗余时间配置是列车晚点恢复时间长短的重要影响因素。本文基于高速铁路列车运行实绩数据,以初始晚点时间、站停冗余时间和区间冗余时间为变量,使用多层感知器(MLP)和循环神经网络(RNN)建立了列车晚点恢复时间预测模型,并采用基于运行图历史数据的冗余时间近似统计方法来提高统计精度,降低了运行图参数数据采集的工作量成本。基于广深港铁路12个月列车运行实绩数据进行了列车晚点恢复时间预测试验,结果表明允许误差为1 min时,MLP模型预测精度为91.6%;允许误差为3 min时, RNN模型表现更好,预测精度在95%以上。
Initial train delay and redundant operation map configuration are two important factors affecting the length of the train recovery time.This study investigates the influencing factors of the train delayrecovery based on high-speed railway train operation records.The initial late time,station stop redundancy time,and interval redundancy time are three important factors affecting the train delay recovery.Based on the high-speed railway train operation data,this study uses multi-layer perceptron(MLP)and a cyclic neural network(RNN)to establish a train delay recovery time prediction model with the initial late time,station stop redundancy time,and interval redundancy time as the variables.The redundant time approximation statistical method based on the historical data of a running graph is adopted to improve the statistical precision and reduce the workload cost of the data collection of the running graph parameters.The train delay recovery time prediction test is performed based on the 12 months train operation datas of the Guangzhou-Shenzhen-HongKong railway.The results show that the prediction accuracy of the MLP model is 91.6%when the allowable error is 1 min.The RNN model is more accurate when the allowable error is 3 min.In conclusion,the prediction accuracy is above 95%.
作者
胡雨欣
彭其渊
鲁工圆
李力
HU Yu-xin;PENG Qi-yuan;LU Gong-yuan;LI Li(School of Transportation and Logistics,Southwest Jiaotong University,Chengdu 611756,China;National United Engineering Laboratory of Integrated and Intelligent Transportation,Chengdu 611756,China)
出处
《交通运输工程与信息学报》
2020年第2期93-102,共10页
Journal of Transportation Engineering and Information
基金
国家重点研发计划(2017YFB1200701)
国家自然科学基金项目(U1834209)。
